• 한국정보통신학회:학술대회논문집
• /
• 한국정보통신학회 2022년도 추계학술대회
• /
• pp.453-456
• /
• 2022

전력망의 역할은 안정적인 전력공급이 최우선이다. 예고 없는 불안정한 상황에 대한 여러 가지 대비에 대한 방안이 필요하다. 기상 데이터를 활용하여 탐구적 데이터 분석을 통한 피처 간의 관계를 파악하여 머신러닝 기반의 에너지 생산 예측 모형을 모델링한다. 본 연구에서는 주성분분석을 사용하여 에너지 생산 예측 시 영향을 미치는 피처를 추출하였으며 머신러닝 모델에 적용함으로써 예측 신뢰도를 높였다. 제안한 모형을 사용하여 특정 기간을 대상으로 생산 에너지를 예측하고 해당 시점의 실제 생산 값과 비교함으로써 주성분분석을 적용한 에너지 생산 예측에 대한 성능을 확인하였다.

• 한국지진공학회:학술대회논문집
• /
• 한국지진공학회 2002년도 추계 학술발표회 논문집
• /
• pp.28-34
• /
• 2002

The b-value in the magnitude-frequency relationship logN(m) = $\alpha$ - bmwhere N(m) is the number of earthquakes exceeding magnitude m, is important seismicity parameter In hazard analysis. Estimation of the b-value for earthquake data observed on KSRS array network is done employing the maximum likelihood technique. Assuming the whole Korea Peninsula as a single seismic source area, the b-value is computed at 0.9. The estimation for KMA earthquake data is also similar to that. Since estimate is a function of minimum magnitude, we can inspect the completeness of earthquake catalog in the fitting process of b-value. KSRS and KMA data lists are probably incomplete for magnitudes less than 2.0 and 3.0, respectively. Examples from probabilistic seismic hazard assessment calculated for a range of b-value show that the small change of b-value has seriously effect on the prediction of ground motion.

• 반도체디스플레이기술학회지
• /
• 제19권4호
• /
• pp.126-129
• /
• 2020

We introduce a lifetime assessment technique using deep learning algorithm with complex electrical parameters such as resistivity, permittivity, impedance parameters as integrated indicators for predicting the degradation of the organic molecules. The evaluation system consists of fully automated in-situ measurement system and multiple layer perceptron learning system with five hidden layers and 1011 perceptra in each layer. Prediction accuracies are calculated and compared depending on the physical feature, learning hyperparameters. 62.5% of full time-series data are used for training and its prediction accuracy is estimated as r-square value of 0.99. Remaining 37.5% of the data are used for testing with prediction accuracy of 0.95. With k-fold cross-validation, the stability to the instantaneous changes in the measured data is also improved.

• 한국철도학회:학술대회논문집
• /
• 한국철도학회 2010년도 춘계학술대회 논문집
• /
• pp.261-267
• /
• 2010

Present maintenance of a high speed railway is after the fack maintenance that executes a task when measured value goes over threshold value except some planned maintenance. It is difficult from efficient management of maintenance human resource and equipment commitment because it is difficult to predict quantity of maintenance targets. Corrective maintenance is pushed back on the repair priority of other target to need repair and it is exceeded repair cost potentially. For safety and dependable track management because track deterioration prediction is linked directly with track's life and safety of train service, it is very important that track management be based on preventive maintenance. In this study, we propose statistics model of track quality to use track inspection data and forecast model for track deterioration prediction.

• Steel and Composite Structures
• /
• 제36권5호
• /
• pp.493-506
• /
• 2020

The load-slip relationship of the shear connection is an important parameter in design and analysis of composite structures. In this paper, a load-slip curve prediction method of the shear connection based on the artificial neural networks (ANNs) is proposed. The factors which are significantly related to the structural and deformation performance of the connection are selected, and the shear stiffness of shear connections and the transverse coordinate slip value of the load-slip curve are taken as the input parameters of the network. Load values corresponding to the slip values are used as the output parameter. A twolayer hidden layer network with 15 nodes and 10 nodes is designed. The test data of two different forms of shear connections, the stud shear connection and the perforated shear connection with flange heads, are collected from the previous literatures, and the data of six specimens are selected as the two prediction data sets, while the data of other specimens are used to train the neural networks. Two trained networks are used to predict the load-slip curves of their corresponding prediction data sets, and the ratio method is used to study the proximity between the prediction loads and the test loads. Results show that the load-slip curves predicted by the networks agree well with the test curves.

• Korean Journal of Radiology
• /
• 제20권4호
• /
• pp.589-598
• /
• 2019

Objective: To evaluate whether data acquired from perfusion computed tomography (PCT) parameters can aid in the prediction of treatment outcome after palliative chemotherapy in patients with unresectable advanced gastric cancer (AGC). Materials and Methods: Twenty-one patients with unresectable AGCs, who underwent both PCT and palliative chemotherapy, were prospectively included. Treatment response was assessed according to Response Evaluation Criteria in Solid Tumors version 1.1 (i.e., patients who achieved complete or partial response were classified as responders). The relationship between tumor response and PCT parameters was evaluated using the Mann-Whitney test and receiver operating characteristic analysis. One-year survival was estimated using the Kaplan-Meier method. Results: After chemotherapy, six patients exhibited partial response and were allocated to the responder group while the remaining 15 patients were allocated to the non-responder group. Permeability surface (PS) value was shown to be significantly different between the responder and non-responder groups (51.0 mL/100 g/min vs. 23.4 mL/100 g/min, respectively; p = 0.002), whereas other PCT parameters did not demonstrate a significant difference. The area under the curve for prediction in responders was 0.911 (p = 0.004) for PS value, with a sensitivity of 100% (6/6) and specificity of 80% (12/15) at a cut-off value of 29.7 mL/100 g/min. One-year survival in nine patients with PS value > 29.7 mL/100 g/min was 66.7%, which was significantly higher than that in the 12 patients (33.3%) with PS value ≤ 29.7 mL/100 g/min (p = 0.019). Conclusion: Perfusion parameter data acquired from PCT demonstrated predictive value for treatment outcome after palliative chemotherapy, reflected by the significantly higher PS value in the responder group compared with the non-responder group.

• 한국IT서비스학회지
• /
• 제21권3호
• /
• pp.63-72
• /
• 2022

The extinction crisis of local cities, caused by a population density increase phenomenon in capital regions, directly causes the increase of vacant houses in local cities. According to population and housing census, Gunsan-si has continuously shown increasing trend of vacant houses during 2015 to 2019. In particular, since Gunsan-si is the city which suffers from doughnut effect and industrial decline, problems regrading to vacant house seems to exacerbate. This study aims to provide a foundation of a system which can predict and deal with the building that has high risk of becoming vacant house through implementing a data driven vacant house prediction machine learning model. Methodologically, this study analyzes three types of machine learning model by differing the data components. First model is trained based on building register, individual declared land value, house price and socioeconomic data and second model is trained with the same data as first model but with additional POI(Point of Interest) data. Finally, third model is trained with same data as the second model but with excluding water usage and electricity usage data. As a result, second model shows the best performance based on F1-score. Random Forest, Gradient Boosting Machine, XGBoost and LightGBM which are tree ensemble series, show the best performance as a whole. Additionally, the complexity of the model can be reduced through eliminating independent variables that have correlation coefficient between the variables and vacant house status lower than the 0.1 based on absolute value. Finally, this study suggests XGBoost and LightGBM based machine learning model, which can handle missing values, as final vacant house prediction model.

• Journal of the Korean Data and Information Science Society
• /
• 제24권4호
• /
• pp.857-865
• /
• 2013

집중 호우로 인한 피해가 증가하면서 다양한 기법들을 이용하여 강우량 예측에 대한 관심이 높아졌다. 최근에는 극단분포를 활용하여 강우량을 예측하려는 시도가 늘고 있다. 본 연구에서는 일반화 극단 분포를 활용하여 실제 서울시의 1973년부터 2010년까지 7월달의 사후예측분포를 생성하고, 수치적인 계산을 위해서 MCMC (Markov chain Monte Carlo)알고리즘을 활용하였다. 이 연구를 통해서 사후예측분포의 점추정값들을 비교하였고 2011년 7월달의 자료와 비교해 봤을 때 집중 호우의 확률이 증가한 것을 알 수 있었다.

• Journal of the Korean Data and Information Science Society
• /
• 제15권2호
• /
• pp.331-337
• /
• 2004

In the analysis of massive volume data, a neural network model is a useful tool. To implement the Neural network model, it is important to select initial value. Since the initial values are generally used as random value in the neural network, the convergent performance and the prediction rate of model are not stable. To overcome the drawback a possible method use samples randomly selected from the whole data set. That is, coefficients estimated by logistic regression based on the samples are the initial values.

• Geomechanics and Engineering
• /
• 제13권6호
• /
• pp.997-1010
• /
• 2017

The black box model is a relatively new option for nonlinear dynamic system identification. It can be used for prediction problems just based on analyzing the input and output data without considering the changes of the internal structure. In this paper, a black box model was presented to solve unconstrained overlying strata movement problems in coal mine production. Based on the black box theory, the overlying strata regional system was viewed as a "black box", and the black box model on overburden strata movement was established. Then, the rock mechanical properties and the mining thickness and mined-out section area were selected as the subject and object respectively, and the influences of coal mining on the overburden regional system were discussed. Finally, a corrected method for height prediction of the fractured zone was obtained. According to actual mine geological conditions, the measured geological data were introduced into the black box model of overlying strata movement for height calculation, and the fractured zone height was determined as 40.36 m, which was comparable to the actual height value (43.91 m) of the fractured zone detected by Double-block Leak Hunting in Drill. By comparing the calculation result and actual surface subsidence value, it can be concluded that the proposed model is adaptable for height prediction of the fractured zone.